Introduction

About 14% of the U.S. population is Latino [1] and most (60%) are Mexican-Americans. Nationally, colorectal cancer (CRC) is the second-most common cancer to be diagnosed among Latinos (11% of the cancers in Latino men and 9% of the cancers in Latino women), after prostate cancer in men and breast cancer in women. In general, increases in CRC screening over time have been hypothesized to be due to greater public awareness of its importance, the adoption of CRC screening as a benefit covered by medical insurance in the private and public sectors, and the implementation of screening programs in various U.S. states [2]. Although Latinos have lower rates of CRC compared to whites, they are diagnosed at a later stage and thus have a lower probability of survival following diagnosis compared to whites [3]. Further, based on the California Cancer Registry, from 1988 to 2002, there was a 26% decrease in invasive CRC incidence among non-Latino whites compared to a 8% decline among Latinos (http://ccrcal.org/Cancer05/ColonRectum05.pdf). Californian Latinos have lower rates of colorectal (CRC) screening compared to non-Latino whites [4] which may account in part for the disparities in colorectal incidence. Understanding why Mexican-Americans have lower rates of CRC screening is important in order to reduce healthcare disparities, a major goal of the Healthy People 2010 national cancer screening objectives [5].

The present study investigated the role that acculturation plays in affecting whether California adults of Mexican origin (Mexicans and Mexican-Americans; hereafter referred to as Mexican-Americans) have CRC screening. CRC screening is comprised of either a fecal occult blood test (FOBT), lower endoscopic procedures, or both FOBT and endoscopy. National cancer screening guidelines [6] recommend that average-risk adults aged 50 years and older have an annual, flexible sigmoidoscopy every 5 years, a combination of FOBT and flexible sigmoidoscopy every 5 years, a double-contrast barium enema every 5 years, or colonoscopy every 10 years. Any positive test results following FOBT, sigmoidoscopy, or enema should be followed by a colonoscopy. Although colonoscopy has been considered the gold standard for CRC screening [7], it is expensive; FOBT plus sigmoidoscopy may be more cost effective in saving lives, and thus more commonly recommended [8].

The purpose of this study was to examine the relationship of acculturation on CRC screening rates among Californian Mexican-Americans. Analyses were adjusted for healthcare access, health status, and socioeconomic status as determined by review of the literature regarding cancer screening behaviors. It was hypothesized that after adjusting the covariates, acculturation would be significantly related to obtaining CRC screening among Mexican-Americans. Analyses were stratified by gender to investigate whether acculturation affects CRC screening rates differentially among Mexican-American men and women. Analyses were delimited to those aged 50 and older, as this is the age targeted for CRC screening guidelines. The analyses did not include those who had been told they had colon cancer, thus focusing on routine CRC screening.

Methods

Participants

A total of 603 Mexican-American men and 893 Mexican-American women aged 50 and older who had not been told they had colon cancer were identified from the 2005 California Health Interview Survey (CHIS) and investigated. The CHIS is a random-digit dial telephone survey of households in California. CHIS interviewed one adult (aged 18 years or older) from each household. The CHIS utilized several techniques to achieve statistically robust samples of ethnic populations.

Measures

CHIS Items

In the 2005 CHIS, two sets of items concerning colorectal screening practices were asked of those aged 40 and older (although for the purpose of analyses, only data for those aged 50 or older were analyzed). The first set of two items relevant to the analyses concerned lower endoscopy (sigmoidoscopy and colonoscopy) and covered whether respondents had ever had the test (yes–no), and whether a doctor recommended endoscopy in the past year (yes–no). The second set of two items concerned at-home FOBT (whether they ever had the test and whether the FOBT was recommended by a doctor in the past year). If needed, during the interview respondents were given a description of each exam in order to clarify each procedure. If requested by the interviewee, the CHIS survey was administered in Spanish by a bilingual interviewer. The survey specifically referred to the at-home FOBT as this test is the only FOBT test demonstrated to be effective in clinical trials [9].

Acculturation Score

An acculturation score was developed by principal components analysis, using the continuous scores automatically generated by the Statistical Package for the Social Sciences (SPSS; Version 14). The acculturation score was based on a unidimensional model [10], emphasized degree of exposure to the dominant culture, and used items available in the 2005 CHIS. The score was composed of seven status variables that measure English language use and proficiency, nativity and citizenship, and years living in the U.S. Scores were scaled to a t-score distribution with a mean of 50 and standard deviation of 10; higher scores indicated higher acculturation.

The reliability and validity of the acculturation score were examined. Cronbach’s alpha for the acculturation score was 0.91; indicating a high level of within-scale inter-item consistency reliability. Evidence of construct validity was observed in the present study through correlations with variables hypothesized to be related to acculturation. Very highly significant differences (P < 0.001) were found between lower- and higher-acculturated Mexican-Americans in the expected direction on income, education, and healthcare access (see Table 1).

Table 1 Descriptive statistics by gender and acculturation status among Mexican-Americans from the 2005 California Health Interview Surveya
Full size table

Covariates

Logistic regression to predict CRC screening included three covariates in addition to acculturation. Three key covariates were identified for inclusion in the logistic regression model based on a review by Hiatt and colleagues [11] and a broad literature review [12–16]. Based on Hiatt review of 65 journal articles concerning predictors of cancer screening behaviors, the authors identified, in addition to attitudinal factors such as acculturation, the following predictors: socioeconomic status, health status, and access to the healthcare system. In the present analysis, socio-demographic status included poverty level and educational level. Health-related factors included age and health status. Access to the health care system included having health insurance, having a usual source of healthcare, seeing a doctor in the past year, and getting a recommendation in the past year to get either an FOBT and/or endoscopy. As the variables were measured on different scales, these were converted to z-scores and summed to create composite scores measuring the three covariates. Higher scores indicated higher education, higher income, higher age, poorer health status, and higher acculturation level, having health insurance, a usual source of healthcare, seeing a doctor in the past year, or having a doctor who recommended CRC screening.

Analysis

Descriptive statistics were obtained for the total, gender, acculturation, and gender by acculturation groups. Differences between groups were analyzed by Chi-square and regression analyses. STATA Version 9 was used for descriptive and logistic regression analyses with an alpha level set at 0.05. Point estimates and logistic regression results took into account complex sampling design using replicate weights to obtain variance estimates based on the jackknife method.

Two logistic regression models were analyzed: one to predict ever having both FOBT and endoscopy (vs. having one or neither test), the other to predict having neither test (vs. having both or either test). The predictor of interest was acculturation status, entered as a three-group dummy variable with the low acculturation group as the reference category. Regression analyses were stratified by gender.

In order to examine effect sizes associated with acculturation in logistic regression models, three descriptive contrast groups were created (low, intermediate, and high) based on three equally sized groups. The three acculturation groups may be described in the following ways. The majority of those in the high acculturation group spoke only English, they were U.S. citizens, were born in the U.S., were interviewed in English, spoke only English, and spoke primarily English in the home. Those at the intermediate level of acculturation indicated that they spoke English very well, were naturalized citizens, born outside the U.S., were interviewed in English, and spoke both Spanish and English in the home. The majority of those in the low acculturation group reported that they didn’t speak English very well, were non-citizens, were born outside of the U.S., were interviewed in Spanish, did not speak English at all or not well, and spoke primarily Spanish in the home. Most respondents had lived in the U.S. for 15 years or more, regardless of acculturation status.

Descriptive analyses (reported in Table 1) and logistic regression analyses (reported in Table 2) were based on Mexican-Americans, and included 893 women and 603 men. For descriptive analyses, a binary acculturation score was created based on a median split of the continuous principal component score which was separately determined by gender.

Table 2 Logistic regression model to predict colorectal cancer screeninga among Mexican-Americans by gender from the 2005 California Health Interview Survey
Full size table

Results

Descriptive Statistics

Table 1 reports descriptive statistics of Mexican-American study participants. The average age of the Mexican-American CHIS 2005 sample under examination was approximately 62 years (standard error 0.3), 74% of whom were above the federal poverty level. A total of 47% had a high school diploma, most were covered by health insurance in the previous year (79%), and had a usual source of healthcare (90%). Overall, about half (55%) were in good, very good, or excellent health. Approximately 18% said that only English was spoken at home; 5% said they had difficulty understanding their doctor at their last doctor visit. Of those (n = 80), 82% said they had difficulty understanding the doctor due to language and 66% said they needed another person to help them understand the doctor.

Women were more likely to report having a poorer health status, a usual source of healthcare, and to have visited a doctor in the past year than men (P < 0.01). Men were more likely to live above the federal poverty level and have obtained a high school diploma (P < 0.05). Compared to lower acculturated participants, higher acculturated respondents were more educated, were more likely to live above the poverty level, have better health status, be covered by health insurance, have a usual source of healthcare, to have seen a doctor in the previous year, speak only English in the home, to not have difficulty understanding the doctor due to speaking another language (P < 0.01). Comparable acculturation-based results were obtained when analyzing men and women separately.

Screening Rates

Higher acculturated Mexican-Americans (see Table 1) as compared to lower acculturated Mexican-Americans were more like to have received a recommendation from their doctor to have either an endoscopy or FOBT (P < 0.001). Higher acculturated Mexican-Americans were more likely to have both endoscopy and FOBT (P < 0.001) compared to the lower acculturated. The lower acculturated were more likely to have had neither endoscopy nor FOBT (P < 0.001) compared to higher acculturated. Among men there was no relationship between acculturation and having FOBT only or endoscopy only (P > 0.10). Women were more likely than men to have had a recommendation from their doctor to have an FOBT in the past year. Higher acculturated men were more likely to receive a recommendation to have an FOBT or an endoscopy compared to lower acculturated men (P < 0.05).

Logistic Regression

Both Endoscopy and FOBT

For Mexican-American men, socioeconomic, health status, and healthcare access factors were marginally or significantly related to having both procedures (P < 0.10). Compared to low acculturated men, high acculturated men were 3.37 times more likely (P = 0.05) to have both CRC screening tests; there was no difference between intermediate and low acculturated men.

Among women, there was a significant relationship between healthcare access and having both screening tests (P < 0.001). After adjustment for covariates, there was a significant effect for acculturation on having both CRC screening tests: compared to low acculturated women, intermediate acculturated women were 1.91 times (P = 0.040) and high acculturated women were 4.08 times (P < 0.001) more likely to have had both CRC screening tests.

Neither Endoscopy Nor FOBT

Mexican-American men of higher socioeconomic status, better health status, and greater healthcare access were less likely to have neither endoscopy nor FOBT (P < 0.01). Lower-acculturated Mexican-American men were about two times as likely to have neither test. Among Mexican-American women, healthcare access was related to having neither test (P < 0.001) but health status and socioeconomic status were not. Low compared to high acculturated women were 2.86 times as likely to have neither CRC screening test (P = 0.001).

Discussion

Colorectal cancer screening among adults aged 50 and older was evaluated as one of the most cost effective preventive tests from among 25 other exams and services [17]. An estimated 50–60% of CRC deaths might be prevented if all U.S. adults aged 50 years or older were routinely screened [18]. The efficacy of FOBT, sigmoidoscopy, and colonoscopy has been estimated as 38%, 50%, and 70%, respectively. Thus, FOBT would lead to a 38% reduction in CRC mortality when used with 100% compliance [19]. According to the American Cancer Society, survival from CRC is about 90% when the cancer is diagnosed early highlighting the importance of early screening and detection of this cancer.

To date, there has been no examination among Mexican-Americans, the largest Latino subgroup in the U.S., of the effect of acculturation on CRC screening rates. The present investigation demonstrated that in California, compared to their lower-acculturated counterparts, higher acculturated Mexican-Americans were 3–4 times more likely to have had both colorectal screening tests and about half as likely to have had neither test. These results could not be explained by socioeconomic, health, or healthcare access status. Not receiving CRC screening may be due to several factors: (1) it is less likely to be offered at the place where lower-acculturated Mexican-Americans receive medical care; (2) their medical care providers are less likely to offer them the test; (3) lower-acculturated Mexican-Americans are more likely to decline the test when it is offered; (4) it is discussed but the medical care providers and patients have difficulty communicating due to language barriers; or (5) lower-acculturated Mexican-Americans are less likely to seek medical care and are subject to fewer outreach efforts.

Implications for Practice

According to some researchers [20], lower rates of CRC screening among lower-acculturated Mexican-Americans may be described as due to the patient or provider barriers. Patient-centered attributes include lack of knowledge about screening, not understanding the role of screening when no problem is present, being embarrassed about the procedure, having cultural attitudes such as fatalism, and not comprehending the English language [20]. Provider-organization attributes include prejudice and racism of clinical staff, lack of cultural competence when discussing colorectal screening with lower-acculturated patients, inability to speak Spanish or provide Spanish-language materials, and not providing CRC screening tests at convenient hours or locations [20].

Mandelblatt and colleagues [20] suggest that communication is the key component that links patients and providers in limiting CRC screening disparities. While lower-acculturated Mexican-American patients may lack knowledge about the importance of CRC screening and a provider may not understand the attitudes about CRC screening held by their Mexican-American patients, communication is the bridge that can help to eliminate disparities in CRC screening. At its most basic level, speaking the same language determines the depth and accuracy at which such health-related communication may occur. The key distinguishing characteristic between higher and lower acculturated Mexican-American patients is English language proficiency. Limited English proficiency patients are less likely to report empathy from their providers, be able to establish a sense of rapport, get the health-related information they need, or collaborate with their doctor to create a treatment plan [21]. Even among insured Latinos, healthcare has been found among the limited English proficient to be poorer, resulting in longer waits, difficulty getting information by phone, and a lack of continuity of care [22]. Results in the present analysis indicate that compared to the higher-acculturated, lower-acculturated Mexican-American women were more likely to have difficulty understanding their doctor due to language and that they needed someone to help understand (P < 0.001).

A recent meta-analysis of the effectiveness of cancer screening services [23] indicated that organizational as opposed to individual interventions were more effective in increasing use of preventive services. They identified seven intervention components that might improve preventive services that were targeted at patients (i.e., reminders, financial incentives), providers (i.e., feedback, education), organizations (i.e., organizational change, regulatory intervention), and communities (i.e., media campaign). Organizational change was the most effective intervention method to increase CRC screening rates (adjusted odds ratio of 17.6). This included establishing a separate clinic specifically to screen patients and giving non-physician staff responsibility for preventive services. Collaboration and teamwork was identified as an important component in fostering the effectiveness of such organizational changes. In addition, provider education (adjusted odds ratio of 3.0) was important: CRC screening guidelines and other information relevant to CRC screening in the Latino community should be disseminated to providers. There are several resources available that can assist providers, including the Centers for Disease Control and Prevention site on CRC (http://www.cdc.gov/cancer/colorectal/) and health disparities (http://www.cdc.gov/omhd/ Topic/HealthDisparities.html); the Cancer Research and Prevention Foundation (http://www.preventcancer.org/); Redes en Acción (Networks in Action) of the NCI-funded National Latino Cancer Research Network (http://www.redesenaccion.org/); the Latino-based health communication outreach organization, Prevención, Inc. (http://www.prevencion.org/), and the American Cancer Society’s Cancer Facts and Figures for Hispanics/Latinos (http://www.cancer.org/docroot/STT/stt_0.asp).

Health promotion methods [3, 23, 24] that target lower-acculturated Mexican-Americans on an individual basis should also be intensified. As some lower-acculturated Mexican-Americans may fear deportation, outreach efforts should be considered. These outreach efforts might focus on improving English language acquisition among Latinos, increasing the Latino community’s knowledge about preventive services and why they are important, creating awareness of Latino rights as health consumers, and helping Latinos to navigate the healthcare system. With respect to the latter, the use of universal symbols (or word pictures) in healthcare settings has been hypothesized [25] as one way to help lower-acculturated Latinos locate and use healthcare services. Universal symbols have the advantage over multilingual signage in that they are easier to implement and maintain and are more easily noticed.

Strengths and Limitations

Differences in screening rates [26] have been demonstrated by Latino subgroup, so focusing specifically on the effect of acculturation among Mexican-Americans was a strength of this investigation. The logistic regression model was based on literature review of the most commonly cited covariates and the three predictors were based on those identified by Hiatt and colleagues [11] who reviewed 65 journal articles concerning predictors of cancer screening behaviors. Another strength of the investigation is that the analysis adjusted for complex sampling designs.

A potential bias is that, because the CHIS is a telephone survey, the results may not be generalizable to people without telephones. However, 95% of the U.S. population has a telephone [27] and the 2005 CHIS data were weighted to minimize this source of bias. In addition, the lack of inclusion of the lower acculturated, who would be less likely to have a telephone, would likely increase the observed odds ratios if they were included in the analyses.

The measure of acculturation was based on conventional acculturation variables [28, 29] such as nativity, citizenship, time spent in the U.S., and facility with and use of the English language. Such measures have been criticized [30] because they are based on proxy rather than direct indicators of acculturation, such as cultural attachment and do not consider ethnic identity, such as an individual’s identification with the Mexican culture. However, the current measure of acculturation had high internal consistency reliability (α = 0.91) even though it included several distinct measures of acculturation-related variables. Further, strong evidence of construct validity was observed as highly significant differences were found between lower- and higher-acculturated Mexican-Americans in the expected direction on differences in income, education, and healthcare access. Lastly, status indicators are objective or observable, easily gathered and are typically the only indicators available in population-based surveys, they are often completed in a more reliable manner than attitudinal measures, and language proficiency has been demonstrated to be a reliable measure of acculturation [29]. Although this may be the first study to relate acculturation to CRC screening rates in a probability sample of Californian Mexican-Americans, similar research should be conducted in a national survey of Mexican-Americans, perhaps using direct measures of bidimensional acculturation according to the Berry model of the acculturation process [31].

Summary

In summary, differences in CRC screening were associated with acculturation in Californian Mexican-Americans. If the preferred colonoscopy screening was increased by 10% and a 70% CRC screening efficacy rate was applied [19], improvements in screening rates could potentially save over 100 lives per year among Californian Latinos aged 45 and older. This represents a 30% increase in the number of lives saved based on the current level of screening using all methods, including FOBT.

One way to decrease the disparity effects of acculturation at the individual level is to provide resources that foster English proficiency as well as health literacy, such as free classes offered at the community level. The effects of acculturation may be diminished by providing low-cost screening clinics in Latino neighborhoods that are staffed with Spanish-speaking Latinos [23]. In other clinical settings universal signage should be used and culturally competent Spanish-speaking clinical providers who are trained in CRC screening procedures should be employed. Reminders concerning screening should be used in clinic settings as this has been found to be an effective way to increase their use [23]. As lower-acculturated Mexican-Americans are less likely to visit a medical clinic, outreach in the form of Spanish language ads on radio and TV might be effective in increasing screening rates. More research is needed on the effect of acculturation on medical outcomes, such as CRC incidence, as paradoxically, although higher acculturated Latinos are more likely to have better access to medical treatment, they are also more likely to engage in poorer health behaviors, such as increased fat intake and decreased physical activity [10].